I have an interesting question about generating ACS tables for a county's unincorporated area. Counties in Colorado have to produce housing tables (such as B25032, Housing tenure by units in structure) for these areas.
In my mind, the most straightforward way to do this is to calculate the residual, i.e., the difference between the categorical totals for the county and the sum of the categorical totals for the municipalities in the county. The works for the estimates, but I’m not sure how to calculate the margin of error in this case.
Imagine a county with two municipalities. The first step would be to calculate the number of housing units in the incorporated areas of the county by aggregating the municipalities, and the second step would be to calculate the residual (which would be the value in the unincorporated area) by calculating the difference between the county total and the aggregated municipal total. For the first step, I would normally use the sum of squares method to estimate the margin of error. I’m not sure what is appropriate for the second step. Chapter 8 of the ACS Handbook has examples for aggregating geographies, but it does not address the case of calculating a residual.
Does anyone have a suggestion for calculating the margin of error for a residual? Is another approach more appropriate?
TIA
AB
In general you can take any linear combination of table cells within or across tables and compute the square of the MoE as the sum of the scaled MoE s for the individual estimates which can be within or…
In general you can take any linear combination of table cells within or across tables and compute the square of the MoE as the sum of the scaled MoE s for the individual estimates which can be within or across tables.
for a sum with + or - signs MoE of combination= square root ( sum MoE ^2)
You do exactly what you said and the take the sum of the square of the sub geography MoE's then take the square root
https://www.census.gov/content/dam/Census/library/publications/2020/acs/acs_general_handbook_2020_ch08.pdf
Remember the "chapter 8" formulas are approximations.
Dave